hmac sha384/sha512 fails test vectors

[iwade]

BPO	6473
Nosy	@amauryfa
Files	test_vectors.py: codification of test vectors from draft rfc.

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = None
closed_at = <Date 2009-07-17.07:11:59.603>
created_at = <Date 2009-07-13.03:56:49.127>
labels = ['type-bug', 'library']
title = 'hmac sha384/sha512 fails test vectors'
updated_at = <Date 2009-07-17.07:11:59.601>
user = 'https://bugs.python.org/iwade'

bugs.python.org fields:

activity = <Date 2009-07-17.07:11:59.601>
actor = 'iwade'
assignee = 'none'
closed = True
closed_date = <Date 2009-07-17.07:11:59.603>
closer = 'iwade'
components = ['Library (Lib)']
creation = <Date 2009-07-13.03:56:49.127>
creator = 'iwade'
dependencies = []
files = ['14492']
hgrepos = []
issue_num = 6473
keywords = []
message_count = 4.0
messages = ['90467', '90579', '90592', '90607']
nosy_count = 3.0
nosy_names = ['amaury.forgeotdarc', 'hagen', 'iwade']
pr_nums = []
priority = 'critical'
resolution = None
stage = 'needs patch'
status = 'closed'
superseder = None
type = 'behavior'
url = 'https://bugs.python.org/issue6473'
versions = ['Python 2.6', 'Python 3.1', 'Python 2.7', 'Python 3.2']

[iwade]

Test vectors are in the following draft rfc:
http://tools.ietf.org/html/draft-nystrom-smime-hmac-sha

The problem is that hmac.py has a hard-coded block size of 64, while
SHA-384 and SHA-512 have a 128-byte block size.

Suggested fix is either:

a/ have the various hashlib libraries export block size (like they
currently do for digest_size).

b/ parameterize blocksize to the constructor so that users can override.

c/ I have no third suggestion.

I have made available a codified version of the test vectors for your
convenience.

[hagen]

Seems like this has already been fixed as bpo-1385.

[amauryfa]

Indeed, the provided test file passes on all python versions I have.
Iain, does this script fail on some version?

[iwade]

d'oh, I should have checked HEAD before submitting the bug.

I am running 2.5.1 on OSX, the fix seems to be in 2.5.2 and above.

Thanks, and sorry for wasting your time.